Generalizing Multi-Context Systems for Reactive Stream Reasoning Applications * 1 Introduction

Stefan Ellmauthaler
unpublished
In the field of artificial intelligence (AI), the subdomain of knowledge representation (KR) has the aim to represent, integrate, and exchange knowledge in order to do some reasoning about the given information. During the last decades many different KR-languages were proposed for a variety of certain applications with specific needs. The concept of a managed Multi-Context System (mMCS) was introduced to provide adequate formal tools to interchange and integrate knowledge between different
more » ... proaches. Another arising field of interest in computer science is the design of online applications, which react directly to (possibly infinite) streams of information. This paper presents a genuine approach to generalize mMCS for online applications with continuous streams of information. Our major goal is to find a good tradeoff between expressiveness and computational complexity. Research in the field of knowledge representation has originated a large variety of formats and languages. To use those formal concepts a wealth of tools have emerged (e.g. databases, ontologies, triple-stores, modal logics, temporal logics, nonmonotonic logics, logic programs under nonmonotonic answer set semantics,. . .). Those tools were designed for specific needs of certain applications in mind. With the idea of a "connected world", nowadays we do not intend to divide information over different applications. It is desirable to have all information available for every application if need be. To express all of this knowledge, represented in specifically tailored languages, in a universal language would be too hard to achieve from the point of view of complexity as well as the troubles arising from the translation of the representations. A second issue in current knowledge representation, which is already addressed in different fields of knowledge representation (e.g. stream data processing and querying [10, 9], stream reasoning with answer set programming [6], forgetting in general [8, 5]), is the lack of online usage of KR tools and formalisms. Most of the approaches only assume one-shot computations, which is triggered by a user. This may be a specific request in the form of a query to a computer. In practice there are many applications where knowledge is provided in a constant flow of information and it is desired to reason over this knowledge in a continuous manner. * This research has been funded by DFG (project FOR 1513)
fatcat:ohw6ksi6mfcohmpwx3omrzqjii